Nutritive, Post-ingestive Signals Are the Primary Regulators of AgRP Neuron Activity  Zhenwei Su, Amber L. Alhadeff, J. Nicholas Betley  Cell Reports  Volume 21, Issue 10, Pages 2724-2736 (December 2017) DOI: 10.1016/j.celrep.2017.11.036 Copyright © 2017 The Author(s) Terms and Conditions

Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions

Figure 1 Nutrients Are Required for the Sustained Reduction of AgRP Neuron Activity (A) Configuration for monitoring calcium dynamics in AgRP neurons. Scale bar, 400 μm. (B) Dual-wavelength fiber photometry (FP) setup used to record calcium-dependent fluorescence (excited at 490 nm) and calcium-independent fluorescence (excited at 405 nm). (C) Food-restricted mice were given two trials with calorie-free gel (CFG) followed by two trials with caloric gel (CG) during FP recordings on 4 consecutive days. Average ΔF/F of GCaMP6s signals from each trial (n = 9 mice/trial) are displayed. Signals are aligned to the first contact with gel at time 0. Green indicates the 490-nm signal; purple indicates the 405-nm control signal. Darker lines represent means, and lighter shaded areas represent SEMs. (D) Heatmaps reporting ΔF/F of the 490-nm signal of the recordings in individual mice in (C). (E) Top: mean ΔF/F of the 490-nm signal from 0 to 400 s among all trials in (C). Bottom: mean ΔF/F of the 405-nm signal from 0 to 400 s among all trials in (C). (F) Mean ΔF/F of the 490-nm signal from trials 1 and 2 displayed in 10-s bins. (G) Mean ΔF/F of the 490-nm signal from trials 2 and 3 displayed in 20-s bins. (H) Mean ΔF/F of the 490-nm signal from trials 3 and 4 displayed in 20-s bins. (I) Percent ΔF/F of the 490-nm signal at first contact with gel relative to maximum ΔF/F in trials 3 and 4 in (C). Values are means ± SEMs. ns, p > 0.05; t tests and post hoc comparisons: ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ANOVA interaction: ∞p < 0.05; ANOVA main effect of group: ☼☼p < 0.01; ☼☼☼p < 0.001. Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions

Figure 2 Nutrients Train the Sensory Regulation of AgRP Neurons in a Single Trial (A) Food-restricted mice were given two trials with CG followed by two trials with CFG during FP recordings. Average ΔF/F of GCaMP6s signals from each trial (n = 8 mice per trial) are displayed. Individual signals were aligned to the first contact with gel at time 0. Green indicates the 490-nm signal; purple indicates the 405-nm control signal. Darker lines represent means, and lighter shaded areas represent SEMs. (B) Heatmaps reporting ΔF/F of the 490-nm signals in individual mice in (A). (C) Mean ΔF/F of the 490-nm signal from trials 1 and 2 displayed in 10-s bins. (D) Percent ΔF/F of the 490-nm signal at first contact with gel relative to maximum ΔF/F in trials 1 and 2 in (A). (E) Mean ΔF/F of the 490-nm signal from trials 2 and 3 displayed in 20-s bins. (F) Mean ΔF/F of the 490-nm signal from trials 3 and 4 displayed in 20-s bins. (G) Mean ΔF/F of the 490-nm signal from 0 to 400 s among all trials in (A). (H) Maximum ΔF/F of the 490-nm signal among all trials in (A). (I) Mean ΔF/F of the 405-nm signal from 0 to 400 s among all trials in (A). (J) Gel intake among trials in (A). Values are means ± SEMs. ns, p > 0.05; t tests and post hoc comparisons: ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ANOVA interaction: ∞∞∞p < 0.001; ANOVA main effect of group: ☼p < 0.05; ☼☼p < 0.01. Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions

Figure 3 GI Nutrients Reduce AgRP Neuron Activity in a Calorie-Dependent Manner (A) Experimental procedure for gastric infusion through implanted catheters during FP recordings. (B) Food-restricted mice were infused with water or 1 kcal of Ensure (n = 8 mice) before chow intake was measured. (C) Average ΔF/F of GCaMP6s signals in mice infused with water (n = 8 mice), hypertonic saline (1.8%, n = 9 mice), sucralose (1.6%, n = 5 mice), methylcellulose (1%, n = 8 mice), or Ensure (1 kcal, n = 9 mice). Individual signals were aligned to the start of infusion at time 0, and infusion period is indicated by gray shading. Green indicated the 490-nm signal; purple indicates the 405-nm control signal. Darker lines represent means, and lighter shaded areas represent SEMs. (D) Top: mean ΔF/F of the 490-nm signal from 0 to 30 min for each infusate. Bottom: mean ΔF/F of the 405-nm signal from 0 to 30 min for each infusate. (E) Maximum ΔF/F of the 490-nm signal for each infusate. (F) Heatmaps reporting ΔF/F of the 490-nm signal for individual mice infused with either 1/3 or 1 kcal of Ensure (n = 9 mice). (G) Mean ΔF/F of the 490-nm signal in 3-min bins from mice infused with 1/3 or 1 kcal of Ensure in (F). Values are means ± SEMs. ns, p > 0.05; t tests and post hoc comparisons: ∗∗∗p < 0.001; ANOVA main effect of group: ☼☼☼p < 0.001. Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions

Figure 4 AgRP Neuron Response Is Proportional to Caloric Content of Food (A) Average ΔF/F of GCaMP6s signals in food-restricted mice given 1/3 or 1 kcal of chow (n = 7 mice). Individual signals are aligned to the delivery of chow at time 0. Green indicates the 490-nm signal; purple indicates the 405-nm control signal. Darker lines represent means, and lighter shaded areas represent SEMs. (B) Heatmaps reporting ΔF/F of the 490-nm signal during chow intake in individual mice. (C) Mean ΔF/F of the 490-nm signal in 3-min bins in mice given 1/3 or 1 kcal of chow in (A). (D) Average ΔF/F of GCaMP6s signals in food-restricted mice given 1/3 or 1 kcal of peanut butter (PB) (n = 7 mice). Individual signals are aligned to the delivery of PB at time 0. (E) Heatmaps reporting ΔF/F of the 490-nm signal during PB consumption in individual mice. (F) Mean ΔF/F of the 490-nm signal in 3-min bins in mice given 1/3 or 1 kcal of PB in (D). (G) Average ΔF/F of GCaMP6s signals in food-restricted mice given 1/3 or 1 kcal of Ensure (n = 7 mice). Individual signals are aligned to the delivery of Ensure at time 0. (H) Heatmaps reporting ΔF/F of the 490-nm signal during Ensure consumption in individual mice. (I) ΔF/F of the 490-nm signal in 3-min bins in mice given 1/3 or 1 kcal of Ensure in (G). (J) Mean ΔF/F of the 490-nm signal from 0 to 30 min in mice given 1/3 kcal of chow, PB, or Ensure. (K) Maximum ΔF/F of the 490-nm signal in mice given 1/3 kcal of chow, PB, or Ensure. (L) Mean ΔF/F of the 490-nm signal from 0 to 30 min in mice given 1 kcal of chow, PB, or Ensure. (M) Maximum ΔF/F of the 490-nm signal in mice given 1 kcal of chow, PB, or Ensure. Values are means ± SEMs. ns, p > 0.05; post hoc comparisons: ∗p < 0.05; ANOVA main effect of group: ☼p < 0.05. Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions

Figure 5 Individual Macronutrients Reduce AgRP Neuron Activity (A) Average ΔF/F of GCaMP6s signals in mice infused with saline (0.9%, n = 9 mice), 1/3 kcal of glucose (n = 8 mice), 1/3 kcal of lipids (n = 6 mice), or 1/3 kcal of amino acids (AAs, n = 7 mice). Individual signals are aligned to the start of infusion at time 0. Green indicates the 490-nm signal; purple indicates the 405-nm control signal. Darker lines represent means, and lighter shaded areas represent SEMs. (B) Top: mean ΔF/F of the 490-nm signal from 0 to 30 min for each infusate. Bottom: mean ΔF/F of the 405-nm signal from 0 to 30 min for each infusate. (C) Maximum ΔF/F of the 490-nm signal for each infusate. (D) Mean ΔF/F of the 490-nm signal in mice infused with 1/3 or 2/3 kcal of glucose in 3-min bins. (E) Mean ΔF/F of the 490-nm signal in mice infused with 1/3 or 1 kcal of lipids in 3-min bins. (F) Mean ΔF/F of the 490-nm signal in mice infused with 1/3 or 1 kcal of AAs in 3-min bins. Values are means ± SEMs. Post hoc comparisons: ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ANOVA main effect of group: ☼p < 0.05; ☼☼p < 0.01; ☼☼☼p < 0.001. Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions

Figure 6 Satiation Signals Reduce AgRP Neuron Activity (A) Diagram showing release sites along the GI tract for eight satiation signals. (B) Average ΔF/F of GCaMP6s signals in mice injected with saline (0.9%, n = 10 mice), cocktail (n = 9 mice), 1/3 dose cocktail (n = 6 mice), LiCl (n = 5 mice), or lipopolysaccharide (LPS, n = 6 mice). Individual signals were aligned to the finish of injection at time 0. Green indicates the 490-nm signal; purple indicates the 405-nm control signal. Darker lines represent means, and lighter shaded areas represent SEMs. (C) Top: mean ΔF/F of the 490-nm signal from 0 to 30 min for each substance injected. Bottom: mean ΔF/F of the 405-nm signal from 0 to 30 min for each substance injected. (D) Maximum ΔF/F of the 490-nm signal for each substance injected. (E) Saccharin solution intake before and 24 hr after conditioned taste avoidance pairing in mice injected with saline (n = 6 mice), cocktail (n = 7 mice), or LiCl (n = 7 mice). (F) Total distance traveled after injection of saline or cocktail (n = 10 mice). (G) Schematic for channelrhodopsin-2 (ChR2)-mediated in vivo photostimulation of AgRP neurons. (H) Saline or peptide was injected i.p. before 1 hr chow re-feeding in food-restricted mice with or without photostimulation. (I) Chow intake in mice with or without cocktail and AgRP neuron stimulation (n = 10 mice). Values are means ± SEMs. ns p > 0.05, t tests and post hoc comparisons: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. Values are means ± SEMs. ns, p > 0.05; post hoc comparisons: ∗∗p < 0.01; ∗∗∗p < 0.001; ANOVA main effect of group: ☼☼p < 0.01. ENT, enterostatin; GLP1, glucagon-like peptide-1; GRP, gastrin releasing peptide; OBE, obestatin; OXY, oxyntomodulin. Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions

Figure 7 CCK, PYY, and Amylin Synergistically Reduce AgRP Neuron Activity (A) Average ΔF/F of GCaMP6s signals in food-restricted mice injected with 30 μg/kg CCK (n = 7 mice), 100 μg/kg PYY (n = 6 mice), or 300 μg/kg amylin (n = 6 mice). Individual signals are aligned to the finish of injection at time 0. Green indicates the 490-nm signal; purple indicates the 405-nm control signal. Darker lines represent means, and lighter shaded areas represent SEMs. (B) Top: mean ΔF/F of the 490-nm signal from 0 to 30 min for each substance injected. Bottom: mean ΔF/F of the 405-nm signal from 0 to 30 min for each substance injected. (C) Mean ΔF/F of the 490-nm signal in 3-min bins from 0 to 30 min among groups injected with cocktail, CCK, PYY, or amylin. (D) Average ΔF/F of GCaMP6s signals in food-restricted mice injected with 3 μg/kg CCK (n = 7 mice); 10 μg/kg PYY (n = 4 mice); 10 μg/kg amylin (n = 4 mice); a cocktail with CCK, PYY, and amylin (C-P-A: 3 μg/kg CCK, 10 μg/kg PYY, and 10 μg/kg amylin; n = 8 mice); or a cocktail with the other 5 peptides (10 μg/kg each of GLP1, GRP, ENT, OBE, and OXY; n = 8 mice). (E) Mean ΔF/F of the 490-nm signal among mice injected with C-P-A, CCK, PYY, or amylin in 3-min bins. (F) Mean ΔF/F of the 490-nm signal from 0 to 30 min among mice injected with saline (n = 10 mice), the other 5 peptides (n = 8 mice), or C-P-A (n = 8 mice). (G) Model for regulation of AgRP neurons. Cell Reports 2017 21, 2724-2736DOI: (10.1016/j.celrep.2017.11.036) Copyright © 2017 The Author(s) Terms and Conditions